en.Wedoany.com Reported - Huawei Cloud released four new products at the 2026 Huawei Cloud INSPIRE Conference: the AICS Lingqu Intelligent Computing Cluster, the AMS Agentic Memory Storage Solution, the CCE VolcanoNext Integrated Intelligent Computing Scheduling Engine, and the AgentSphere Runtime Environment. Zhou Yuefeng, Huawei Board Member and CEO of Huawei Cloud, stated at the conference that the era of Agentic AI is driving a fundamental shift in computing paradigms, with cloud computing entering the Token industrialization phase. Post-training is evolving from static parameter tuning to autonomous evolution through Agentic RL, and the underlying infrastructure is accelerating towards integrated intelligent computing. Based on this, Huawei Cloud formally proposed the new Agentic Infra paradigm, whose core includes "Efficient Token Factory + Continuous Learning + Integrated Intelligent Computing Scheduling + Secure Autonomy."

The AICS Lingqu Intelligent Computing Cluster, based on the ultra-large bandwidth Lingqu network, supports a cluster scale of up to 100,000 cards, with a total computing power of 200 EFLOPS. This cluster reduces Token generation latency to under 10 milliseconds, achieves a throughput of 5 million Tokens per second per thousand cards, and maintains an online service availability of 99.95%. The AMS Agentic Memory Storage Solution leverages NPU-direct CMS (Context Memory Storage) hardware to create a PB-level massive memory space, supporting KV Cache hierarchical pooling, which reduces inference costs while enabling day-level long-duration tasks. The CCE VolcanoNext Integrated Intelligent Computing Scheduling Engine achieves load scheduling innovation through "shared training and inference pools + fragment consolidation," improving resource utilization by over 30%. AgentSphere, as a secure and autonomous Agent runtime environment, provides ultra-fast elasticity and intent-based proactive protection. With its lightweight sandbox technology, it achieves 100-millisecond-level ultra-fast startup and a batch creation capability of hundreds of thousands per minute, facilitating the secure and efficient large-scale migration of intelligent agents to the cloud.

Zhou Yuefeng stated that only by optimizing Token costs, integrated scheduling, RL (Reinforcement Learning) efficiency, and secure autonomy to the highest degree can one seize the initiative in developing infrastructure for the Agentic AI era. Huawei Cloud will leverage collaborative innovation across software, hardware, and chips to build an enterprise-grade "silicon-based black soil" for Agentic AI, thereby leading the new Agentic Infra paradigm and laying the foundation for accelerating the deployment of applications in the Agentic era.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com